Let RS RS Find approximations for EG and Var using Taylor expansions of For any xy the bivariate 64257rst order Taylor expansion about is xy remainder Let EXEY The simplest approximation for XY is then XY The approximation for XY ID: 33549
Download Pdf The PPT/PDF document "Approximations for Mean and Variance of ..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Forf(R;S)=R=S,thederivativesaref00RR(R;S)=0;f00RS(R;S)=S2;andf00SS(R;S)=2R S3.Specically,when=(R;S),wehavef()=R=S;f00RR()=0;f00RS()=1 (S)2;andf00SS()=2R (S)3.ThenanimprovedapproximationofE(R=S)isE(R=S)E(f(R;S))R SCov(R;S) (S)2+Var(S)R (S)3(11)Bythedenitionofvariance,thevarianceoff(X;Y)isVar(f(X;Y))=En[f(X;Y)E(f(X;Y))]2o(12)UsingE(f(X;Y))f()(fromabove)Var(f(X;Y))En[f(X;Y)f()]2o(13)ThenusingtherstorderTaylorexpansionforf(X;Y)expandedaroundVar(f(X;Y))Ehf()+f0x()(Xx)+f0y()(Yy)f()i2(14)=Ehf0x()(Xx)+f0y()(Yy))i2(15)=Enf02x()(Xx)2+2f0x()(Xx)f0y()(Yy)+f02y()(Yy)2o(16)=f02x()Var(X)+2f0x()f0y()Cov(X;Y)+f02y()Var(Y)(17)Nowwereturntoourexample:f(R;S)=R=Sexpandedaround=(R;S).Sincef0R=S1;f0S=R S2and=(R;S),wenowhavef02R()=1 (S)2;f0R()f0S()=R (S)3;f02S()=(R)2 (S)4.andsoVar(R=S)1 (S)2Var(R)+2R (S)3Cov(R;S)+(R)2 (S)4Var(S)(18)=(R)2 (S)2"Var(R) (R)22Cov(R;S) RS+Var(S) (S)2#(19)=(R)2 (S)2"2R (R)22Cov(R;S) RS+2S (S)2#(20)Reference:Kendall'sAdvancedTheoryofStatistics,Arnold,London,1998,6thEdition,Vol-ume1,byStuart&Ord,p.351.Reference:SurvivalModelsandDataAnalysis,JohnWiley&SonsNY,1980,byElandt-JohnsonandJohnson,p.69.